Re: [RFC PATCH 04/10] perf workqueue: add threadpool execute and wait functions

[Namhyung Kim]

On Tue, Jul 13, 2021 at 5:11 AM Riccardo Mancini <rickyman7@xxxxxxxxx> wrote:
>
> This patch adds:
>  - execute_in_threadpool: assigns a task to the threads to execute
>    asynchronously.
>  - wait_threadpool: waits for the task to complete on all threads.
> Furthermore, testing for these new functions is added.
>
> This patch completes the threadpool.
>
> Signed-off-by: Riccardo Mancini <rickyman7@xxxxxxxxx>
> ---
>  tools/perf/tests/workqueue.c           |  86 ++++++++++++++++++++-
>  tools/perf/util/workqueue/threadpool.c | 103 +++++++++++++++++++++++++
>  tools/perf/util/workqueue/threadpool.h |   5 ++
>  3 files changed, 193 insertions(+), 1 deletion(-)
>
> diff --git a/tools/perf/tests/workqueue.c b/tools/perf/tests/workqueue.c
> index be377e9897bab4e9..3c64db8203556847 100644
> --- a/tools/perf/tests/workqueue.c
> +++ b/tools/perf/tests/workqueue.c
> @@ -1,13 +1,59 @@
>  // SPDX-License-Identifier: GPL-2.0
> +#include <stdlib.h>
>  #include <linux/kernel.h>
> +#include <linux/zalloc.h>
>  #include "tests.h"
>  #include "util/debug.h"
>  #include "util/workqueue/threadpool.h"
>
> +#define DUMMY_FACTOR 100000
> +#define N_DUMMY_WORK_SIZES 7
> +
>  struct threadpool_test_args_t {
>         int pool_size;
>  };
>
> +struct test_task {
> +       struct task_struct task;
> +       int n_threads;
> +       int *array;
> +};
> +
> +/**
> + * dummy_work - calculates DUMMY_FACTOR * (idx % N_DUMMY_WORK_SIZES) inefficiently
> + *
> + * This function uses modulus to create work items of different sizes.
> + */
> +static void dummy_work(int idx)
> +{
> +       int prod = 0;

I'm not sure but having 'volatile' would prevent some kind of
possible compiler optimizations..

> +       int k = idx % N_DUMMY_WORK_SIZES;
> +       int i, j;
> +
> +       for (i = 0; i < DUMMY_FACTOR; i++)
> +               for (j = 0; j < k; j++)
> +                       prod ++;
> +
> +       pr_debug3("dummy: %d * %d = %d\n", DUMMY_FACTOR, k, prod);
> +}
> +
> +static void test_task_fn1(int tidx, struct task_struct *task)
> +{
> +       struct test_task *mtask = container_of(task, struct test_task, task);
> +
> +       dummy_work(tidx);
> +       mtask->array[tidx] = tidx+1;
> +}
> +
> +static void test_task_fn2(int tidx, struct task_struct *task)
> +{
> +       struct test_task *mtask = container_of(task, struct test_task, task);
> +
> +       dummy_work(tidx);
> +       mtask->array[tidx] = tidx*2;
> +}
> +
> +
>  static int __threadpool__prepare(struct threadpool_struct **pool, int pool_size)
>  {
>         int ret;
> @@ -38,21 +84,59 @@ static int __threadpool__teardown(struct threadpool_struct *pool)
>         return 0;
>  }
>
> +static int __threadpool__exec_wait(struct threadpool_struct *pool,
> +                               struct task_struct *task)
> +{
> +       int ret;
> +
> +       ret = execute_in_threadpool(pool, task);
> +       TEST_ASSERT_VAL("threadpool execute failure", ret == 0);
> +       TEST_ASSERT_VAL("threadpool is not executing", threadpool_is_busy(pool));
> +
> +       ret = wait_threadpool(pool);
> +       TEST_ASSERT_VAL("threadpool wait failure", ret == 0);
> +       TEST_ASSERT_VAL("waited threadpool is not ready", threadpool_is_ready(pool));
> +
> +       return 0;
> +}
>
>  static int __test__threadpool(void *_args)
>  {
>         struct threadpool_test_args_t *args = _args;
>         struct threadpool_struct *pool;
> -       int ret;
> +       int ret, i;
> +       struct test_task task;
> +
> +       task.task.fn = test_task_fn1;
> +       task.n_threads = args->pool_size;
> +       task.array = calloc(args->pool_size, sizeof(*task.array));

Need to check the return value.

>
>         ret = __threadpool__prepare(&pool, args->pool_size);
>         if (ret)
>                 return ret;
>
> +       ret = __threadpool__exec_wait(pool, &task.task);
> +       if (ret)
> +               return ret;
> +
> +       for (i = 0; i < args->pool_size; i++)
> +               TEST_ASSERT_VAL("failed array check (1)", task.array[i] == i+1);
> +
> +       task.task.fn = test_task_fn2;
> +
> +       ret = __threadpool__exec_wait(pool, &task.task);
> +       if (ret)
> +               return ret;
> +
> +       for (i = 0; i < args->pool_size; i++)
> +               TEST_ASSERT_VAL("failed array check (2)", task.array[i] == 2*i);
> +
>         ret = __threadpool__teardown(pool);
>         if (ret)
>                 return ret;
>
> +       free(task.array);

All previous returns will leak it.

Thanks,
Namhyung